[1M3-GS-10] AI application: Materials

[1M3-GS-10-03] Proposing Method for Automotive Structural Component Cross-Sectional Designs through Diffusion Model

〇Tsuyoshi Nishihara¹, Eri Kaiki¹, Kaori Suzuki¹, Keita Ohmine¹, Toshiaki Yokoi², Shugo Nakamura², Masahiro Nakamoto² (1. Mazda Motor Corporation, 2. DENTSU SOKEN INC.)

Keywords:Surrogate model, Diffusion model, Simulation, indust

In the development of automotive structural components, there is a demand to ensure high energy absorption performance while designing lightweight structures within a short timeframe. Surrogate models are effective means for efficiently conducting structural investigations; however, structural generation is often parametric, limiting the freedom of shape variation. In this study, we apply a diffusion model to propose cross-sectional shapes that meet the target energy absorption performance. By employing AI to suggest component structures, innovation in automotive parts design is anticipated.